feat(environment): stream lines in ReadFileTool to prevent OOM on large files

Aperta
#7,131 3 commenti 0 reazioni 1 assegnatario Vedi su GitHub

@sanketpatil06 ci sta già lavorando.

Dal 16/9/2026.

Valutazione

Questa issue non è ancora stata valutata.

Descrizione

needs review tools

Required Information

Is your feature request related to a specific problem?

Currently, ReadFileTool loads an entire file into memory as bytes via await self._environment.read_file(path) and then splits lines across the whole payload before slicing the requested range (start_line to end_line).

For very large files (hundreds of megabytes or gigabytes), reading the entire file into memory risks high memory usage or Out-Of-Memory (OOM) crashes in the agent environment, even when the caller only requests a small range of lines. There is an explicit TODO in src/google/adk/tools/environment/_read_file_tool.py:
# TODO: Avoid loading the entire file into memory to prevent OOM on large files.

Describe the Solution You'd Like
  1. Add a read_file_lines(path, start_line, end_line) method to BaseEnvironment with a backward-compatible default fallback that calls read_file.
  2. Implement an optimized, streaming read_file_lines on LocalEnvironment that iterates through the file lazily line-by-line in $O(1)$ memory, only collecting the requested range.
  3. Update ReadFileTool.run_async to call read_file_lines, eliminating the need to buffer the entire file into memory.
Impact on your work

Improves stability and performance of file inspection tools in agent environments handling large log files, datasets, and source files.

Willingness to contribute

Yes, I have an implementation ready with unit tests and pre-commit checks passing.

Recommended Information

Proposed API / Implementation
# BaseEnvironment
async def read_file_lines(
    self,
    path: str | Path,
    start_line: int = 1,
    end_line: int | None = None,
) -> tuple[list[bytes], int]: ...

# LocalEnvironment
@staticmethod
def _sync_read_lines(
    path: Path, start_line: int, end_line: int | None = None
) -> tuple[list[bytes], int]:
  selected_lines: list[bytes] = []
  total_lines = 0
  with open(path, 'rb') as f:
    for line in f:
      total_lines += 1
      if start_line <= total_lines and (
          end_line is None or total_lines <= end_line
      ):
        selected_lines.append(line)
  return selected_lines, total_lines
Lingua principale
Python
Stelle
21.6k
Fork
4k
Merge medio
13h 49m
PR unite (30g)
10

Guida per i contributori

Apri la guida per i contributori

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di google/adk-python

Tutte le issue di google/adk-python

Issue simili

Altre issue su Python

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.